1999 Alfred Wegener Conference Abstract (Van Ark)

Van Ark, E., F. Marton, S. Stein, C. Bina, and D. C. Rubie, Persistence of metastable olivine in detached slabs as a possible cause of deep earthquakes, Terra Nostra, Abstracts of the Alfred Wegener Conference on Processes and Consequences of Deep Subduction, Verbania, Italy, 99/7, 106, 1999.

Certain deep earthquakes occur in patterns which suggest the presence of a slab detached from near-surface manifestations of subduction. One example is the deep seismicity beneath the North Fiji Basin, thought to take place in a slab of former Pacific plate subducted at the Vitiaz trench and then detached during a late Miocene arc reversal (8-10 Ma). It has been suggested that such deep earthquakes result from transformational faulting in metastable olivine regions which persist for some time after the slab detachment. We examine that possibility using thermo-kinetic modeling of the metastable olivine region within a detached slab to trace the evolution of the metastable region as it gradually transforms to the stable mineral phase. This evolution can be characterized by the area of the initial metastable region at the time of detachment and the total persistence time of the region. We first investigate the influence of various lithospheric thermal parameters and subduction parameters on the initial metastable area and persistence time. We then use plausible conditions for subduction at the Vitiaz trench in the model, and compare the resulting metastable wedge with the seismogenic Vitiaz region. If all the seismic zone is required to be suitable for transformational faulting, a 10,000 km2 cross-sectional area of metastable olivine is implied by the seismicity of the past 50 years. According to the model, 8 to 10 My after the detachment of a Vitiaz-type slab the area of the metastable region would be less than 10 percent of that area. This observation is reminiscent of the previously noted difficulty that large deep earthquakes seem to have occurred on faults larger than expected for either the metastable region or the low-temperature region required by other thermally-controlled process which might cause deep earthquakes. Hence for transformational faulting to be a viable cause, either the cold area must be larger than expected, perhaps due to complicated slab geometry, or faulting nucleated by a transformational mechanism must propagate beyond the metastable region.